www.minitab.

com


MINITAB ASSISTANT WHITE PAPER
This is one of a series of papers that explains the research conducted by Minitab
statisticians to develop the methods and data checks used in the Assistant in
Minitab 16 Statistical Software.
GAGE R&R STUDY (CROSSED)
Overview
Measurement system studies are performed in virtually every type of
manufacturing industry to properly monitor and improve a production process. In a
typical measurement system study, a gage is used to obtain repeated
measurements on selected parts by several operators. Two components of
measurement system variability are frequently generated in such studies:
repeatability and reproducibility. Repeatability represents the variability when the
gage is used to measure the same part by the same operator. Reproducibility refers
to the variability from different operators measuring the same part. Thus,
measurement system studies are often referred to as gage repeatability and
reproducibility studies, or gage R&R studies.
The primary purpose of a gage study is to determine how much variation in the
data is due to the measurement system, and whether the measurement system is
capable of assessing process performance. For detailed discussions on
measurement system studies, refer to the MSA manual (2003), Montgomery and
Runger (1993), and Burdick, Borror, and Montgomery (2005).
The Gage R&R Study (Crossed) command in the Assistant is designed to analyze
data from typical measurement system studies. It adopts the most common
approach of fitting the measurement data with an ANOVA model and estimates
different sources of variation in the measurement system using the variance
components in the model.
If you use the typical guidelines for how much data to collect for gage R&R studies,
the variance components may not be precisely estimated (Montgomery and Runger,
1993a, 1993b; Vardeman and VanValkenburg, 1999). The Assistant indicates
whether the number of parts and the number of operators are less than certain
values, which may affect the precision of the part-to-part and operator variation
estimates. We conducted simulations to identify the number of parts, operators,
and replicates that are needed to obtain precise estimates.





Gage R&R Study (Crossed) 2
Using our simulation results and widely accepted practices in measurement system
analysis, we developed the following data checks for Gage R&R Study (Crossed).
The Assistant automatically performs these data checks and reports the findings in
the Report Card.
- Amount of Data
- Xbar Chart (repeatability)
- R Chart (reproducibility)
In this paper, we investigate how these data checks relate to measurement system
analysis in practice and we describe how we established the guidelines for each
data check.
Data checks
Amount of data
Typically, guidelines for gage R&R studies recommend using 10 parts, 2 or 3
operators, and 2 or 3 replicates (AIAG, 2003; Raffaldi and Ramsier, 2000; Tsai,
1988). However, the recommended sample size is not large enough to estimate
part-to-part variation with good precision and, therefore, may not provide a good
basis for assessing whether or not to use a particular gage (Montgomery and
Runger, 1993a, 1993b; Vardeman and VanValkenburg, 1999).
To establish guidelines for the appropriate amount of data, we focused on how
many parts should be evaluated to obtain estimates of part-to-part variation with
different levels of precision. We also evaluated how many operators should be used
to obtain a precise estimate of measurement variation. Finally, we investigated the
number of observations required to obtain gage repeatability estimates with
different precisions.
Number of parts to estimate part-to-part variation
with different levels of precision
Objective
We wanted to determine how many parts should be evaluated to obtain estimates
of part-to-part variation with different levels of precision.
Method
We performed a simulation study using 5000 samples. For all samples, we
estimated the standard deviation of the parts and calculated the ratio of the
estimated standard deviation to the true standard deviation. We sorted the ratios





Gage R&R Study (Crossed) 3
from low to high and used the 125
th
and 4875
th
ratios to define the 95% confidence
interval; the 250
th
and 4750
th
ratios define the 90% confidence interval. Using these
confidence intervals, we identified how many parts are needed to estimate part-to-
part variation with different levels of precision.
Results
Based on the simulation study, we concluded the following:
- Using 10 parts, 3 operators, and 2 replicates, the ratio of the 90% confidence
interval over the true standard deviation is about (0.61, 1.37) with 35% to
40% margin of error. At 95% confidence, the interval is about (0.55, 1.45)
with 45% margin of error. Therefore, 10 parts are not enough to produce a
precise estimate for the part-to-part variation component.
- You need approximately 35 parts to have a 90% confidence of estimating the
part-to-part variation within 20% of the true value.
- You need approximately 135 parts to have a 90% confidence of estimating
the part-to-part variation within 10% of the true value.
We also determined that these results apply to acceptable, marginal, and
unacceptable gages.
See Appendix A for a detailed explanation of the simulation and its results.
Number of operators to estimate operator variation
with different levels of precision
Objective
We wanted to determine how many operators should evaluate parts to obtain
operator variation estimates with different levels of precision.
Method
The standard deviation for parts and the standard deviation for operators are both
estimated using the ANOVA model. Therefore, the method used in the simulation
for the number of parts to estimate part-to-part variation also applies to the
number of operators to estimate the variation between operators.
Results
Two or three operators are not enough to provide a precise estimate for
reproducibility. However, the problem is less critical if the magnitude of part-to-part
variation is much larger than the variation among operators, which is a likely
scenario for many applications.





Gage R&R Study (Crossed) 4
See Appendix A for a detailed explanation of the simulation and its results.
Number of observations to estimate repeatability with
different levels of precision
Objective
We wanted to determine how the number of observations affects the estimate of
repeatability and whether 10 parts, 3 operators, and 2 replicates can provide a
reasonably precise estimate for repeatability variation.
Method
The ratio of the estimated repeatability standard deviation over its true value
follows a chi-square distribution. To determine the number of observations needed
to obtain a reasonably precise estimate of repeatability, we calculated the lower
and upper bounds of the ratio associated with 90% probability and graphed the
results.
Results
In a typical gage study (for example, number of parts = 10, number of operators =
3, and number of replicates = 2), the degrees of freedom for error equals 30, which
allows you to have about 90% confidence of estimating the repeatability within
20% of the true value. Under typical settings, the estimate for repeatability is
reasonably precise. See Appendix B for more details.
Overall results
Our studies clearly indicate that the typical settings used in a gage study are not
good enough to provide precise estimates for part-to-part variation and
reproducibility variation, which affect the ratio of the gage variation over the total
process variation, and ultimately the decision about whether the gage is acceptable.
Typically, part-to-part variation is greater than reproducibility variation, and
therefore its precision has a greater impact on whether to accept a gage. However,
in many applications, it may not be feasible to select 35 or more parts and have
multiple operators measure them twice.
Considering the typical gage R&R settings used in practice and our simulation
results, the Assistant uses the following approaches to encourage users to obtain
precise estimates for the variance components:
1. Provide an option in the dialog box to allow users to enter an estimate of
process variation obtained from a large historical data set. In most cases, the
estimate from a large historical data set has better precision than the
estimate from the sample data.





Gage R&R Study (Crossed) 5
2. If the historical estimate is not available, and the number of parts is small,
we display a message to remind users to select more than 10 parts to obtain
more precise estimates.
Based on the amount of data, the Assistant Report Card displays information about
process variation and measurement variation. For example, if you use 10 parts and
3 operators and specify a historical standard deviation, the following data check is
displayed in the Report Card:

Status Condition

To determine if a measurement system is capable of assessing process
performance, you need good estimates of the process variation and the
measurement variation.
- Process variation: Comprised of part-to-part and measurement
variation. It can be estimated from a large sample of historical data,
or from the parts in the study. You entered a historical standard
deviation so both estimates are available. You can compare them to
see how well they agree. Although the number of parts in this study
(10) satisfies the typical requirement of 10, the historical value
should provide a more precise estimate of the process variation.
- Measurement variation: Estimated from the parts, it is broken down
into Reproducibility and Repeatability. The number of parts (10) and
operators (3) meets the typical requirement of 10 parts and 3
operators. This is usually adequate for estimating Repeatability, but
the estimate of Reproducibility is less precise. If the %Process for
Reproducibility estimate is large, you may want to examine the
differences between operators and determine if these differences are
likely to extend to other operators.
Below are all the messages for various configurations of parts, operators, and
replicates.
Process variation
Historical standard deviation (parts < 10)
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You entered a historical standard deviation so both estimates are
available. You can compare them to see how well they agree. Because the
number of parts in this study is small, the historical value should provide a
more precise estimate of the process variation.





Gage R&R Study (Crossed) 6
Historical standard deviation (parts 10, 15)
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You entered a historical standard deviation so both estimates are
available. You can compare them to see how well they agree. Although the
number of parts in this study satisfies the typical requirement of 10, the
historical value should provide a more precise estimate of the process
variation.
Historical standard deviation (parts > 15, < 35)
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You entered a historical standard deviation so both estimates are
available. You can compare them to see how well they agree. The number of
parts in this study is much larger than the typical requirement of 10. If the
selected parts represent typical process variability, this estimate of the
process variation should be much better than if you used 10 parts.
Historical standard deviation (parts 35)
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You entered a historical standard deviation so both estimates are
available. You can compare them to see how well they agree. The number of
parts in this study is much larger than the typical requirement of 10. If the
selected parts represent typical process variability, this estimate of the
process variation should be adequate.
No historical standard deviation (parts < 10):
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You chose to estimate from the parts but have fewer than the
typical requirement of 10. The precision of this estimate may not be
adequate. If the selected parts do not represent typical process variability,
consider entering a historical estimate or using more parts.
No historical standard deviation (parts 10, 15):
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You chose to estimate from the parts. Although the number of
parts satisfies the typical requirement of 10, the estimate may not be
precise. If the selected parts do not represent typical process variability,
consider entering a historical estimate or using more parts.





Gage R&R Study (Crossed) 7
No historical standard deviation (parts > 15, < 35):
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You chose to estimate from the parts. The number of parts is
much larger than the typical requirement of 10. If the selected parts
represent typical process variability, this estimate of the process variation
should be much better than if you used 10 parts
No historical standard deviation (parts 35):
- Process variation: Comprised of part-to-part and measurement variation. It
can be estimated from a large sample of historical data, or from the parts in
the study. You chose to estimate from the parts. The number of parts is
much larger than the typical requirement of 10. If the selected parts
represent typical process variability, this estimate of the process variation
should be adequate.
Measurement variation
Operators 2 or Parts < 10
- Measurement variation: Estimated from the parts, it is broken down into
Reproducibility and Repeatability. The number of parts and operators does
not meet the typical requirement of 10 parts and 3 operators. The estimates
of measurement variation may not be precise. You should view the estimates
as indicating general tendencies, rather than precise results.
Operators 3 and 5 and parts 10
- Measurement variation: Estimated from the parts, it is broken down into
Reproducibility and Repeatability. The number of parts or operators meets
the typical requirement of 10 parts and 3 operators. This is usually adequate
for estimating Repeatability, but the estimate of Reproducibility is less
precise. If the %Process for Reproducibility estimate is large, you may want
to examine the differences between operators and determine if these
differences are likely to extend to other operators.
Operators > 5 and parts 10
- Measurement variation: Estimated from the parts, it is broken down into
Reproducibility and Repeatability. The number of parts or operators meets
the typical requirement of 10 parts and 3 operators, and is usually adequate
for estimating Repeatability. The additional operators improve the precision
of the Reproducibility estimate.





Gage R&R Study (Crossed) 8
Xbar chart
In gage R&R studies, the Xbar chart is a specialized version of an Xbar control
chart. The chart plots the averages of the multiple readings by each operator on
each part, and the control limits are calculated using the repeatability variation. The
repeatability variation from an acceptable gage should be much less than the part-
to-part variation, which is reflected by the variation of the plotted points on the
chart. Therefore, with an acceptable gage, most plotted points should fall outside
the control limits.
Objective
We wanted to develop a guideline on whether the measurement system has a
repeatability issue based on the percentage of points that fall outside the control
limits of the Xbar chart.
Method
We researched the literature to determine how many out-of-control points are
needed to signal a potential problem with repeatability.
Results
The MSA manual (2003) recommends that approximately 50% or more of the
plotted points should be outside the control limits. We adopted the MSA
recommendation to check the Xbar Chart in the Assistant Gage R&R Study output.
Based on this recommendation, the Assistant Report Card for Gage R&R Study
(Crossed) displays the following status indicators for the Xbar chart:

Status Condition

If any points are outside the limits
The control limits are based on Repeatability. Ideally, the variation from
repeated measurements is much less than the variation between parts.
Guidelines suggest that approximately 50% or more should fall outside
the limits. In this study, xx.x% of points are outside.

If no points are outside the limits
The control limits are based on Repeatability. Ideally, the variation from
repeated measurements is much less than the variation between parts.
Guidelines suggest that approximately 50% or more should fall outside
the limits. In this study, no points are outside.





Gage R&R Study (Crossed) 9
R chart
In gage R&R studies, an R chart is used to check the reproducibility variation.
Specifically, you can answer the following questions with an R chart:
- Does each operator measure all the parts consistently? If not, which part is
more difficult to measure consistently?
- Do all operators have similar measurement variation? Do specific operators
measure with significantly greater variation?
Objective
We wanted to help users determine whether their measurement system can
measure parts consistently.
Method
We decided to use the common approach of creating an R chart to display the
variation in measurements made by each operator.
Results
Points on the R chart that are above the control limit indicate that operators did not
consistently measure some of the parts. The Assistant calculates the percent of
points that are above the upper control limit.
The Assistant Report Card for Gage R&R Study (Crossed) displays the following
status indicators for the R chart:

Status Condition

Some points are outside the limits
Each point is the range of the measurements for a part. In this study,
xx.x% of the points are above the upper control limit, indicating parts
were measured inconsistently. Try to understand why the measurements
are inconsistent and determine whether there were any data entry
errors.

If no points are outside the limits
Each point is the range of the measurements for a part. In this study, no
points are above the upper control limit, indicating all parts were
measured with similar consistency.





Gage R&R Study (Crossed) 10
References
Burdick, R.K., Borror, C. M., and Montgomery, D.C. (2005). Design and analysis of
gauge R&R studies: Making decisions with confidence intervals in random and
mixed ANOVA models. Philadelphia, PA: Society for Industrial Applied Mathematics
(SIAM).
Automotive Industry Action Group (AIAG) (2003). Measurement systems analysis
(MSA) manual (3rd edition). Southfield, MI: Chrysler, Ford, General Motors Supplier
Quality Requirements Task Force.
Montgomery, D.C. (2000). Design and analysis of experiments. New York, NY:
Wiley.
Montgomery, D.C., and Runger, G.C. (1993 a). Gage capability and designed
experiments. Part I: Basic methods. Quality Engineering, 6 (1993/1994), 115
135.
Montgomery, D.C., and Runger, G.C. (1993 b). Gage capability analysis and
designed experiments. Part II: Experimental design models and variance
component estimation. Quality Engineering, 6 (1993/1994), 289-305.
Raffaldi, J. and Ramsier, S. (2000). 5 ways to verify your gages. Quality Magazine,
39 (3), 38-42.
Tsai, P. (1988). Variable gage repeatability and reproducibility study using the
analysis of variance method. Quality Engineering, 1(1), 107-115.
Vardeman, S.B. and VanValkenburg, E.S. (1999). Two-way random-effects
analyses and gage R&R studies. Technometrics, 41 (3), 202-211.





Gage R&R Study (Crossed) 11
Appendix A:
Evaluate the effect of parts on part-to-
part variation
Because there is no exact formula to calculate the confidence interval for the part-
to-part standard deviation, we performed a simulation to estimate the interval. To
focus the simulation on how the number of parts affects the precision of the
estimated part-to-part variation, we examined the ratio of the estimated confidence
interval for the standard deviation of the parts over the true standard deviation of
the parts. As the number of parts increases, the interval becomes narrower. We
then identified the number of parts such that the margin of error for the ratio is
10% or 20%. The interval for the 10% margin of error is (0.9, 1.1), and for the
20% margin of error is (0.8, 1.2).
Simulation setup
A gage R&R study assumes that the k
th
measurement of the i
th
part by the j
th

operator, denoted as

, fits the following model:

Where
, and

, and

are independently normally distributed with mean 0, and

variances of

, and

. Here

, and

represent parts, operators,

parts x operators, and error terms.
Let r be the ratio of the total gage standard deviation over the total process
standard deviation. Then,

Typically, the following rule is used to determine whether a measurement system is
acceptable:
r 0.1 (10%): acceptable
0.1 < r 0.3: marginal
0.3 < r: unacceptable
We choose r = 0.1 (acceptable), r = 0.25 (marginal), and r = 0.35 (unacceptable)
to define the three regions. For the purposes of the simulation, we assume that the
repeatability variance equals the reproducibility variance, which gives:





Gage R&R Study (Crossed) 12

We use

and 1,

, and

to generate the
observations, and assume that 3 operators measure each part twice to evaluate
how the number of parts affects the standard deviation of the parts.
These are the simulation steps we followed for each number of parts, r, and

:
1. Generate 5000 samples using the model above.
2. Estimate part standard deviation, and calculate the ratio of the estimated
standard deviation over the true standard deviation for all 5000 samples.
3. Sort the 5000 ratios in increasing order. Of the 5000 sorted ratios, the 125
th

and 4875
th
ratios represent the lower and upper bounds of the interval at the
95% confidence level, and the 250
th
and 4750
th
ratios represent the lower
and upper bounds of the interval at the 90% confidence level.
4. Examine the intervals to identify the number of parts such that the margin of
error is 10% or 20%. The interval for the 10% margin of error is (0.9, 1.1).
The interval for the 20% margin of error is (0.8, 1.2).
Simulation results
The results in Tables 1-6 show the simulation results at each confidence level for
different numbers of parts, with each table corresponding to a specific combination
of values for r and

. Overall, these results show that:

- Using 10 parts, 3 operators, and 2 replicates, the ratio of the 90% confidence
interval over the true standard deviation is about (0.61, 1.37) with 35% to
40% margin of error. At the 95% confidence level, the interval is about
(0.55, 1.45) with 45% margin of error. Therefore, 10 parts are not enough to
produce a precise estimate for the part-to-part variation component.
- You need approximately 35 parts to have a 90% confidence of estimating the
part-to-part variation within 20% of the true value.
- You need approximately 135 parts to have a 90% confidence of estimating
the part-to-part variation within 10% of the true value.
Note that this summary of the results is not specific to a particular combination of r
and

The rows corresponding to the bulleted results above are highlighted in

Tables 1, 2, 3, 4, 5, and 6 below.






Gage R&R Study (Crossed) 13
Table 1: Acceptable gage (r = 0.1),

, true part stdev = 0.014071247

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
3 (0.15295, 1.93755) (0.22195, 1.73365)
5 (0.34415, 1.67035) (0.41861, 1.53873)
10 (0.55003, 1.44244) (0.60944, 1.36992)
15 (0.63295, 1.36927) (0.68721, 1.30294)
20 (0.68532, 1.31187) (0.72950, 1.25701)
25 (0.71230, 1.27621) (0.75578, 1.23251)
30 (0.74135, 1.24229) (0.77645, 1.20841)
35 (0.76543, 1.23033) (0.80066, 1.19706)
50 (0.79544, 1.20337) (0.82636, 1.16595)
100 (0.85528, 1.13696) (0.88063, 1.11635)
135 (0.87686, 1.12093) (0.89448, 1.09760)
140 (0.88241, 1.11884) (0.90130, 1.09974)
Table 2: Acceptable gage (r =0.1),

, true part stdev = 14.071247

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
5 (0.34656, 1.68211) (0.42315, 1.55880)
10 (0.55496, 1.45382) (0.61319, 1.38233)
15 (0.63484, 1.36949) (0.68767, 1.30505)
35 (0.76233, 1.23513) (0.79749, 1.19623)





Gage R&R Study (Crossed) 14
Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
40 (0.77256, 1.21518) (0.81224, 1.18121)
135 (0.88017, 1.12345) (0.89883, 1.10249)
140 (0.88004, 1.11725) (0.89787, 1.09713)
145 (0.88281, 1.11886) (0.89966, 1.09583)
150 (0.88302, 1.11132) (0.90096, 1.09296)
Table 3: Marginal gage (r = 0.25),

, true part stdev = 0.005477225575

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
30 (0.73879, 1.25294) (0.77982, 1.21041)
35 (0.75881, 1.24383) (0.79848, 1.20068)
40 (0.77281, 1.22813) (0.80369, 1.18788)
135 (0.87588, 1.11910) (0.89556, 1.10093)
140 (0.87998, 1.12001) (0.89917, 1.09717)
145 (0.88100, 1.11812) (0.89852, 1.09710)
150 (0.88373, 1.11563) (0.90345, 1.09706)
Table 4: Marginal gage (r = 0.25),

, true part stdev = 5.477225575

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
30 (0.74292, 1.25306) (0.78159, 1.20872)





Gage R&R Study (Crossed) 15
Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
35 (0.76441, 1.24391) (0.79802, 1.20135)
40 (0.77525, 1.21339) (0.80786, 1.17908)
135 (0.87501, 1.11711) (0.89512, 1.09758)
140 (0.87934, 1.11756) (0.89881, 1.09862)
145 (0.88308, 1.11530) (0.90056, 1.09806)
Table 5: Unacceptable gage (r = 0.35),

, true part stdev = 0.00378504

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
30 (0.74313, 1.25135) (0.77427, 1.20568)
35 (0.75409, 1.24332) (0.79444, 1.19855)
40 (0.76582, 1.22289) (0.80599, 1.18615)
135 (0.87641, 1.12043) (0.89507, 1.09820)
140 (0.87635, 1.11539) (0.89651, 1.09368)
145 (0.88339, 1.11815) (0.89772, 1.09591)
Table 6: Unacceptable gage (r = 0.35),

, true part stdev = 3.78504

Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
30 (0.73750, 1.26100) (0.77218, 1.21285)
35 (0.74987, 1.23085) (0.79067, 1.18860)





Gage R&R Study (Crossed) 16
Ratio of estimated confidence interval
for part stdev/true part stdev
Number of parts 95% Confidence 90% Confidence
40 (0.77187, 1.22270) (0.80648, 1.18329)
135 (0.87572, 1.11877) (0.89409, 1.09827)
140 (0.87798, 1.11634) (0.89590, 1.09695)
145 (0.87998, 1.11513) (0.89683, 1.09534)
Number of operators
The standard deviation for parts and the standard deviation for operators are
estimated identically using the ANOVA model. Therefore, the simulation results on
parts also apply to reproducibility variation. Two or three operators are not enough
to provide a precise estimate for reproducibility. However, the problem is less
critical for operators if the magnitude of part-to-part variation is much larger than
the operator variation, which is a likely scenario for many applications.
For example, suppose part-to-part standard deviation is 20 times the operator
standard deviation. The part standard deviation is 20, and operator standard
deviation is 1. Assuming repeatability is the same as reproducibility, then the true
ratio of measurement system variation over the total process variation is:

Now assume the margin of error for estimating the operator standard deviation is
40% (high). That is, the estimated operator standard deviation could be 1.4.
Therefore, the ratio of the measurement system overall the total becomes:

Because this value is less that 0.10, a large reproducibility variation does not affect
gage acceptance if 10% is the cutoff value.
If the operator variation is nearly the same as part variation, you need a large
number of operators to represent the measurement system and to accurately
evaluate the gage.





Gage R&R Study (Crossed) 17
Appendix B:
Estimating repeatability
Calculation setup
Unlike confidence intervals for the part-to-part standard deviation, which are based
on an approximation, the ratio of the estimated repeatability standard deviation
over its true value follows a chi-square distribution. Therefore, we can calculate the
lower and upper bounds of the ratio associated with 90% probability, and then
evaluate how both bounds approach 1 as the number of parts, number of
operators, and the number of replicates increase.
Using the same notation defined in Appendix A, the repeatability variance is
estimated by

) 1 ( / ) (
2
.
2
=

K IJ Y Y S
ij ijk


Then,
2
2
) 1 (
e
S K IJ
o

follows a chi-square distribution with IJ(K-1) degrees of freedom

(df), where I is the number of parts, J is the number of operators, and K is the
number of replicates.

Based on this result, the ratio of the estimated standard deviation over its true
value satisfies the following probability equation:

o
_
o
_
o o
=
|
|
.
|

\
|
s s

1
2
) 2 / ( 1 ,
2
. 2 / ,
df
S
df
y Probabilit
df
e
df
.

where df = IJ(K-1) = number of parts * number of operators * (number of
replicates 1). If the number of replicates equals 2, the degrees of freedom equal
the number of parts times the number of operators.
Using this formula, for each given value of the degrees of freedom, we calculate the
lower and upper bounds of the ratio
e
S
o
at a probability of 90%. We then identify
the degrees of freedom such that the estimated standard deviation is within 10%





Gage R&R Study (Crossed) 18
and 20% of its true value. The corresponding interval is (0.9, 1.1) for the 10%
margin of error, and (0.8, 1.2) for the 20% margin of error.
Calculation results
The graph in Figure 1 shows the lower and upper bounds of the ratio
e
S
o
at 90%
probability versus the degrees of freedom, with the degrees of freedom ranging
from 1 to 200.
200 150 100 50 0
2.0
1.5
1.0
0.5
0.0
Degrees of Freedom
R
a
t
i
o 1.1
1.2
0.8
0.9
Lower
Upper
Variable
Lower and Upper Bounds of the Ratio at 90% Probability

Figure 1: Lower and upper bounds of
e
S
o
at 90% probability versus degrees of
freedom (1 to 200)

Notice that that the interval formed by the lower and upper bounds narrows as the
degrees of freedom increase. The width of the interval decreases dramatically as
the degrees of freedom increase from 1 to 50. We can see this more clearly in the
enlarged graph shown in Figure 2, which displays the results for degrees of freedom
from 1 to 50.






Gage R&R Study (Crossed) 19
50 40 30 20 10 0
2.0
1.5
1.0
0.5
0.0
Degrees of Freedom
R
a
t
i
o 1.1
1.2
0.8
0.9
5 10 20 30 40
0.5
0.7
1.3
1.5
Lower
Upper
Variable
Lower and Upper Bounds of the Ratio at 90% Probability

Figure 2: Lower and upper bounds of
e
S
o
at 90% probability versus degrees of
freedom (1 to 50)
As shown in Figure 2, when the degrees of freedom are less than 10, the interval is
wider than (0.63, 1.35). As the degrees of freedom increase, the interval becomes
narrower, as indicated by the values in Table 7 below.

Table 7: Degrees of freedom and lower and upper bounds at 90% probability

Degrees of
freedom
Interval formed
by lower and
upper bounds
5 (0.48,1.49)
10 (0.63, 1.35)
15 (0.70, 1.29)
20 (0.74, 1.25)
25 (0.76, 1.23)





Gage R&R Study (Crossed) 20
Degrees of
freedom
Interval formed
by lower and
upper bounds
30 (0.79, 1.21)
35 (0.80, 1.19)
40 (0.81, 1.18)

Therefore, at 90% probability, you need about 35 degrees of freedom to obtain a
20% margin of error for the standard deviation estimate of repeatability. Recall that
the degrees of freedom equal the Number of Parts * Number of Operators *
(Number of Replicates 1). Therefore, the typical recommendation of 10 parts, 3
operators, and 2 replicates provides degrees of freedom (30) that are close to this
requirement. To obtain a 10% margin of error at 90% probability, you need about
135 degrees of freedom (see Figure 1).

Minitab, Quality Companion by Minitab, Quality Trainer by Minitab, Quality.
Analysis. Results and the Minitab logo are all registered trademarks of Minitab,
Inc., in the United States and other countries.

2010 Minitab Inc. All rights reserved.